Print this page and work through it on paper. Revealed answers are never printed, so you can check one question and still print a clean sheet.

  1. Show the solution to question 1

    (a) Complete the square in each variable. Group the terms by variable and move the constant across:

    Add , and to both sides to complete the three squares:

    This is the equation of the sphere with centre and radius . The right-hand side came out positive, so it is a genuine sphere and not a point or the empty set.

    (b) Compare a distance with the radius. By the distance formula, . Since , , so the origin is farther from the centre than the radius: it lies outside .

    Second justification: the original left-hand side equals , which is negative inside the sphere, zero on it and positive outside. At the origin it equals , so outside. (Equivalently .)

    (c) Slice with a horizontal plane. Substituting into the completed-square form gives

    This has solutions exactly when , that is , so . At the two extremes and the plane is tangent and the intersection is a single point; at it passes through the centre and cuts a great circle of radius .

    For the equation is : a circle lying in the plane , with centre (directly above ) and radius .

    (d) Read the inequalities geometrically. The first inequality is , the solid ball of radius centred at . The plane passes through the centre, and keeps the half of the ball on one side of it. The region is a solid hemisphere (half-ball) of radius : its flat face is the disc of radius centred at in the plane , and its curved face bulges toward positive .

  2. Show the solution to question 2

    (a) Normalise. . The unit vector in the direction of is

    A vector of length in the opposite direction is . Check: , and it is a negative multiple of , so it points the opposite way.

    (b) Solve for the unknown vector exactly as for a number. From , .

    Check: .

    (c) Magnitude is not additive. , so . Meanwhile , so . Since , : the sum of the lengths is larger.

    This is the triangle inequality , with equality only when the vectors point the same way, i.e. one is a positive scalar multiple of the other. Here and are not parallel at all (the ratio of first components is , of second components ), so strict inequality was guaranteed before any arithmetic.

    (d) Head equals tail plus displacement. . Since , .

  3. Show the solution to question 3

    (a) Expand both squares with the dot product. Using , distributivity and commutativity,

    and in the same way . Adding, the cross terms cancel and

    Geometrically, and are the two diagonals of the parallelogram with sides and , so the identity says: the sum of the squares of the two diagonals equals the sum of the squares of the four sides.

    (b) Subtract the same two expansions instead of adding. Directly,

    because .

    For a genuine parallelogram and are non-parallel, so both diagonals are nonzero vectors. Nonzero vectors are perpendicular if and only if their dot product is zero, and by the identity that happens if and only if , i.e. . Opposite sides of a parallelogram are equal anyway, so says all four sides are equal: the diagonals are perpendicular exactly when the parallelogram is a rhombus.

    (c) Feed the data into (a) and into the expansion. From the parallelogram identity,

    For the angle, the expansion of gives , so . By the geometric form of the dot product,

    Cross-check with the other expansion: , agreeing with the first calculation.

  4. Show the solution to question 4

    (a) Expand along the first row.

    The middle component carries the minus sign: . Check perpendicularity: and .

    , so the two unit vectors perpendicular to both are . There are exactly two because the perpendicular direction to a plane is unique up to sign.

    (b) Area is the length of the cross product. The parallelogram has area . The triangle with those two sides from the common vertex is half of it: .

    (c) First way: use the product already computed. Since ,

    Second way: the determinant. Expanding along the first row,

    (Or compute and dot with : .)

    The parallelepiped has volume : it is flat, so the three vectors are coplanar. Being coplanar, and with , not parallel, must be a combination of the other two. Solve componentwise:

    Adding the first two gives , so and . The third equation is the check: . Hence ; verify: .

  5. Show the solution to question 5

    (a) Compare directions, then try to solve for a common point. The direction vectors are and . Neither is a scalar multiple of the other (the first components would need the factor , the second components ), so the lines are not parallel.

    If they meet, some and satisfy all three coordinate equations:

    Subtracting the first equation from the second eliminates : , so , and then . Test these values in the third equation: the left side is , the right side is . They disagree, so no common point exists. Not parallel and not intersecting: the lines are skew.

    (b) Distance via a common normal. The lines lie in a pair of parallel planes whose common normal is perpendicular to both directions:

    and any nonzero multiple will do, so take . (Check: , .) Pick on () and on (), so . The distance is the absolute scalar projection of this vector onto :

    (c) A plane through with normal . A plane containing and parallel to must have a normal perpendicular to both and , which is exactly . Through :

    It contains : substituting gives for every . It is parallel to : substituting gives for every , a constant, so never crosses the plane and in fact lies in the parallel plane .

    Second route to (b): the distance from any point of , say , to the plane is

    the same value. (Equivalently the two parallel planes and are apart.)

    (d) Substitute the parametric equations into the plane.

    The point is . Check in the plane: . (A single solution was guaranteed because : the line is not parallel to the plane.)

  6. Show the solution to question 6

    (a) Complete the square in and in . Group and factor the leading coefficient out of the -terms:

    Collecting constants, , and dividing by ,

    Three squared terms, exactly one negative, and on the right: a hyperboloid of one sheet. Its centre is and its axis is the line , , parallel to the -axis — the variable carrying the minus sign.

    (b) Slice. Substitute the constant into the standard form.

    : , an ellipse centred at with semi-axes in the -direction and in the -direction. This is the waist, the smallest horizontal cross-section.

    : , ellipses centred at with semi-axes and — the same shape as the waist, scaled by .

    : , i.e. : a hyperbola in the plane opening in the directions, vertices at and , asymptotes .

    : : a hyperbola in the plane opening in the directions, vertices at and , asymptotes .

    (c) The drawing. Draw the axis as a vertical line through and mark the centre. Draw the waist ellipse at height (width in , in ), and the larger ellipses at heights and (width in , in ). Join them with the two hyperbolic profiles: in the front-to-back plane the outline passes through and and flares outward above and below; in the side plane it passes through and and flares likewise. The surface opens along the -axis in both directions, getting wider without bound as grows (the ellipses at height have semi-axes and ). It is connected because every horizontal plane meets it in a nonempty ellipse: for all , so there is no gap.

    (d) Vary the constant. With constant the completed-square computation gives

    If the right-hand side is : is an elliptic cone with vertex , and the horizontal traces pinch to a point there. If the right-hand side is negative; dividing by it flips the signs to , two negative squared terms: a hyperboloid of two sheets, with no points where . If the right-hand side is positive and the surface is a hyperboloid of one sheet, as in (a). The cone at is the boundary case between the two hyperboloids, and it is their common asymptotic surface.

  7. Show the solution to question 7

    (a) Intersect the component domains; take the limit componentwise. needs ; needs ; needs . The domain is the intersection, .

    For the limit, each component is its own single-variable problem. for , so the first component tends to ; by continuity; by continuity. The limit is .

    (b) Project onto the -plane, then lift. A point on both surfaces satisfies , that is after completing the square. So the projection of onto the -plane is the circle of radius centred at . Parametrise it by , , and let the plane supply the height :

    Verification on the paraboloid: . On the plane: . Both hold for every , and as runs over the projection goes once round its circle, so is traversed once. (Since lies in the plane and projects to a circle, it is an ellipse.)

    The height is , largest when and smallest when . Highest point: , . Lowest point: , . Check: has , and the origin lies on both surfaces trivially.

    (c) Solve and watch for extraneous roots. On , reads . Squaring both sides,

    So or . Squaring may have added solutions, so test each candidate in the original equation. gives or : at , ✓; at , ✗, extraneous (it solves instead). gives : ✓.

    The curve meets the plane at exactly two points: and . Both have , and both satisfy and : and . (A closed curve crossing a plane generically does so an even number of times, which is a comforting check.)

  8. Show the solution to question 8

    (a) Product rule for the dot, then for the cross. By the dot product rule,

    The first term is zero because is perpendicular to both of its factors, in particular to , and the dot product of perpendicular vectors is . For the second, the cross product rule (factors kept in order) gives

    Here , because the cross product of a vector with itself has length (a vector is parallel to itself). What survives is

    (b) Differentiate the squared length. Let . By the dot product rule,

    for every , by hypothesis. A real function with zero derivative on an interval is constant, so for some constant , and therefore is constant. (The hypothesis says the velocity is always tangent to the sphere through the current point; the conclusion is that the point never leaves that sphere.)

    (c) Chain rule on the square root. Where the function is a composition of the differentiable function (for ) with , so by the chain rule and the dot product rule,

    For : . At , , , and . So the distance from the origin is changing at the rate

    while the speed is .

    They differ because the formula just proved is the scalar projection of onto : , where is the angle between the position and the velocity. Only the radial part of the velocity changes the distance from the origin; the part perpendicular to moves the point sideways around the origin without changing how far away it is. Since , the rate of change of distance can never exceed the speed, and indeed because .

  9. Show the solution to question 9

    (a) Integrate componentwise. Each component is an ordinary integral:

    and, integrating by parts with , ,

    The integral is .

    (b) Match all three coordinates with independent parameters. A common point needs and with

    Substituting the first into the second gives , which is : an identity, so the first two equations impose only one condition, . (Both curves project onto the same line in the -plane, which is why.) The third equation decides everything: gives , so , and . There is exactly one solution, hence exactly one meeting point:

    Tangent vectors: so ; so . By the geometric form of the dot product,

    so , an acute angle (the dot product is positive).

    (c) Point plus direction. The tangent line passes through with direction :

    It meets the -plane where : , , giving the point .

    (d) Find a linear relation among the coordinates. On , for every ; on , for every . So both curves lie in the plane , whose normal is . A vector is parallel to the plane exactly when it is perpendicular to : and , as they must be, since the tangent to a curve lying in a plane lies in that plane. (Within the plane each curve is a parabola in the height ; two parabolas of the same width , shifted, cross exactly once, which is the geometric reason behind (b).)

  10. Show the solution to question 10

    (a) Check the equation of the cone. , so every point of the curve satisfies . Since the curve lies on the upper nappe , spiralling outward and upward as increases (the distance from the -axis is , which grows, while the angle in the -plane is ).

    (b) Speed, length, arc length function. Differentiate each component with the product rule:

    Squaring and adding, the mixed terms cancel:

    so . The speed is never zero and is continuous, so the curve is smooth and the arc length formula applies:

    The arc length function from is . Solving, , so for . (Check: gives , matching .)

    (c) Curvature by the cross product formula. Differentiate once more:

    Factor out of each vector and write , for brevity: . Expanding the determinant,

    since . Its squared length is , so . (Check that is perpendicular to : .) Hence

    As , : the spiral's turns get ever larger in radius (the distance from the axis is ), and a bend of larger radius is gentler. The curvature falls off like the reciprocal of that radius, exactly as for circles.

    (d) Dot the tangent with . By the geometric form of the dot product the angle between and satisfies

    independent of . So the tangent makes the same angle with the axis of the cone at every point: the curve climbs the cone at a constant pitch, the way a helix climbs a cylinder. (The same computation with in place of gives and , so the tangent also makes a constant angle with the position vector, with cosine .)

  11. Show the solution to question 11

    (a) Steps 1 and 2 are correct; Steps 3 and 5 are wrong; Step 4 inherits the error of Step 3.

    Step 3 confuses turning per unit time with turning per unit arc length. Curvature is defined as , and converting to the parameter costs a factor of the speed: . The formula is valid only when is arc length, i.e. when the speed is ; here the speed is . The computation of itself is right (differentiating is legitimate precisely because the speed is constant), but the last equality of the step is not.

    Step 5 draws the wrong conclusion from a failed check. The identity has no extra hypothesis: it holds for every smooth motion, because with and perpendicular unit vectors, and Pythagoras does the rest. A check that fails is evidence of an error upstream, never evidence against the theorem. The student computed correctly and should have gone back to look for the mistake.

    (b) Divide by the speed. With and ,

    a constant, as it should be for a helix. Then

    The check now passes: . With the whole acceleration is normal, and indeed points horizontally toward the axis, which is the direction of .

    (c) The cross product route. With and ,

    Its length is . Therefore

    agreeing with (b). (Notice that the student's wrong is exactly with the division by omitted — the same missing factor in a different guise.)

    The osculating circle has radius , which exceeds the cylinder's radius . The horizontal projection of the path is a circle of radius , but the actual path also climbs: the vertical velocity stretches each turn into a longer, gentler arc, so the curve bends less sharply than its shadow does. Only if the climb were removed ( constant) would rise to and fall to .

  12. Show the solution to question 12

    (a) Differentiate the cross product. By the cross product rule (factors kept in order),

    The first term is because any vector is parallel to itself (sine of the angle is ). For the second, substitute the hypothesis and pull the scalar out: , again because parallel vectors have zero cross product. So for all ; each component has zero derivative, hence is constant, and is a constant vector.

    Now is perpendicular to both of its factors, so for every . If this says the position vector always satisfies the single linear equation : the particle stays in the plane through the origin with normal vector . It is the plane spanned by the initial position and initial velocity, and , since .

    (b) Differentiate the two squared lengths. Using and the dot product rule,

    Substitute in the first term and use commutativity of the dot product:

    A real function with zero derivative is constant, so is conserved. (Physically is twice the energy per unit mass of a particle on an ideal spring: kinetic plus potential.)

    (c) Check the example. and . ✓

    Since both and have zero third component, has only a -component, equal to :

    constant, and the plane is , where the motion indeed takes place. For the energy,

    constant. (The path is the ellipse , traversed once per period .)

    (d) Components at . , so the speed is , and . Then

    Check: . ✓ The curvature is (equivalently ), so the osculating circle at has radius . The value is independent of , as it must be: curvature is a property of the ellipse, not of how fast it is traversed.

    Why there: is the derivative of the speed, not the speed itself. Here , which is not constant — it oscillates between and — but at it is at its maximum , and a differentiable function has zero derivative at an interior maximum. So the speed is momentarily not changing at , even though it is about to decrease; all of the acceleration at that instant goes into turning. (At , the end of the long axis, the same computation gives again, , and : the ellipse bends most sharply at the ends of its major axis and least at the ends of its minor axis.)

Full solutions sheet →